Table 1. Comparative classification performance of AV-Net.

Comparison	Cross Validation	Accuracy	F1	IOU
UNet Dice + Focal Loss Pre-trained Weights	Artery	88.054 ± 0.343	77.743 ± 0.673	63.711 ± 0.879
	Vein	88.653 ± 0.704	78.978 ± 0.722	65.354 ± 0.953
	Average	88.353 ± 0.500	78.360 ± 0.675	64.533 ± 0.884
AV-Net Dice Loss Pre-trained Weights	Artery	83.570 ± 0.734	61.769 ± 1.480	44.871 ± 1.495
	Vein	80.125 ± 1.587	63.697 ± 1.401	47.017 ± 1.388
	Average	81.848 ± 1.020	62.733 ± 1.423	45.944 ± 1.422
AV-Net Focal Loss Pre-trained Weights	Artery	81.007 ± 0.404	46.031 ± 1.236	29.980 ± 1.045
	Vein	81.566 ± 0.249	46.137 ± 0.942	30.066 ± 0.801
	Average	81.287 ± 0.307	46.084 ± 0.964	30.023 ± 0.816
AV-Net Dice + Focal Loss Random Weight	Artery	85.957 ± 0.842	73.243 ± 1.377	58.018 ± 1.656
	Vein	85.501 ± 0.582	73.631 ± 1.258	58.412 ± 1.544
	Average	85.729 ± 0.572	73.437 ± 1.304	58.215 ± 1.584
AV-Net Dice + Focal Loss Pre-trained Weights	Artery	86.705 ± 1.087	82.761 ± 1.677	70.658 ± 2.404
	Vein	86.798 ± 1.174	82.850 ± 1.666	70.781 ± 2.399
	Average	86.751 ± 1.126	82.805 ± 1.670	70.719 ± 2.399